AI Research
Track important developments across AI models, agents, research, business, hardware, robotics, regulation, safety and security — with source attribution and links to the original reporting.
Important AI developments right now.
Introducing AlphaProof Code: a breakthrough in verifying complex software at depth
AlphaProof Code agents wrote three million lines of proofs in Lean to verify Signal’s core encryption.
Read original ↗Closing the AI guidance gap in education
New ThinkYoung research and lessons from OpenAI’s EMEA education partnerships on supporting teachers, AI literacy, and independent learning.
Read original ↗Work at the Frontier: AI’s effect on freelance work
New research from OpenAI and Upwork finds shifting demand for freelance work, new skill combinations, and larger budgets and more hours for AI-related projects.
Read original ↗Google’s AlphaGenome 2 zooms in on the cellular roots of disease
Researchers at Google DeepMind have unveiled a new AI model designed to unravel how changes in our DNA affect different cells, giving scientists a more detailed view of a notoriously difficult problem…
Obscured Wildfire Flame Detection by Spatio-temporal Analysis of Smoke Patterns Using Frame-wise Transformers
arXiv:2611.14074v1 Announce Type: cross Abstract: In this paper, we propose a new paradigm in wildfire management for detecting both visible and obscured fire using Deep Learning (DL). The essence of our method…
AI Isn't Getting Dumber, We're Getting Pickier: Rising Standards in Human Evaluation of Large Language Models
arXiv:2611.14333v1 Announce Type: cross Abstract: Users often complain that AI assistants such as ChatGPT are getting worse. Some versions do change, but human evaluation leaves a second possibility untested: the people judging…
DisasterAgent: An Evidence-Driven Agent for Verifiable Disaster Report Generation From Remote Sensing Imagery
arXiv:2611.13776v1 Announce Type: cross Abstract: Remote sensing disaster assessment is moving from simple damage classification toward comprehensive report-level analysis. In this setting, responders need not only damage descriptions but also verifiable evidence…
The Case for AI-Enabled Gate Monitoring as Farm Security in the UK: A PRISMA-Guided Review of the Evidence on Rural Crime, Livestock Predation and Farm Access
arXiv:2611.13933v1 Announce Type: cross Abstract: Farm security in the UK rests on thinner evidence than its cost suggests. We ran a PRISMA guided review: twelve searches of Web of Science and Scopus…
AI Governance Should Prioritize Control and Knowledge Boundaries Over Limiting Intelligence
arXiv:2611.12941v1 Announce Type: new Abstract: The rapid proliferation of autonomously acting AI agents makes the governance of agentic systems urgent. We propose the competence circuit, summarized as IKC + KC + C,…
Graph Visualizations in Prompts for Knowledge Graph Question Answering
arXiv:2611.12938v1 Announce Type: new Abstract: We study when and how multimodal input improves knowledge graph question answering (KGQA) in large vision--language models. Using a controlled evaluation over 1,500 WebQSP and CWQ questions…
Tracing Psychometric Inference in Large Language Models
arXiv:2611.12937v1 Announce Type: new Abstract: Do large language models represent psychological constructs internally, or do they only generate outputs that resemble psychometric structure? We address this question using a cross-persona paradigm across…
LandslideFM: A Foundation Model Towards Unified Landslide Perception and Understanding
arXiv:2611.12936v1 Announce Type: new Abstract: Intelligent landslide detection is essential for disaster prevention. However, existing models primarily focus on isolated visual tasks, require extensive labeled data, and lack deep analytical and scene-understanding…
TACT: Mitigating Overthinking and Overacting in Coding Agents via Activation Steering
arXiv:2605.05980v3 Announce Type: replace Abstract: When language model agents tackle complex software engineering tasks, they often degrade over long trajectories, which we define as *agent drift*. We focus on two recurring failure…
SetPrint: Black-Box LLM Fingerprinting via Setwise Correspondence of Response Changes
arXiv:2611.12950v1 Announce Type: new Abstract: Black-box LLM fingerprinting aims to determine whether a target model was derived from a source model when only the output responses to input queries are observable. One…
Introducing AlphaProof Code: a breakthrough in verifying complex software at depth
AlphaProof Code agents wrote three million lines of proofs in Lean to verify Signal’s core encryption.
Closing the AI guidance gap in education
New ThinkYoung research and lessons from OpenAI’s EMEA education partnerships on supporting teachers, AI literacy, and independent learning.
Work at the Frontier: AI’s effect on freelance work
New research from OpenAI and Upwork finds shifting demand for freelance work, new skill combinations, and larger budgets and more hours for AI-related projects.
MIT-led team receives funding to develop next-generation alternatives to modern weed killers
With up to $53 million from ARPA-H, the researchers will use AI to generate safer herbicides that reduce dependence on imports and improve the security of the U.S. food supply.
Say Shibboleth: Fingerprinting the LLMs and Harnesses Behind Agentic Attacks
arXiv:2611.12386v1 Announce Type: cross Abstract: We study the fingerprinting of agentic cyberattacks. Our objective is to reliably infer information about the attacker's configuration, such as LLM version and harness, from telemetry. To…
HiVe-Mem: Interleaving Horizontal Memory Expansion and Vertical Evidence Composition for Multimodal Agents
arXiv:2611.12872v1 Announce Type: cross Abstract: As multimodal agents accumulate increasingly long interaction histories, effectively retrieving and utilizing past information becomes critical for long-term interaction. Existing memory systems typically rely on semantic retrieval…
What Is the Value of AI-Involved Art? Evidence From a Juried In-Person Exhibition
arXiv:2611.10150v1 Announce Type: new Abstract: What is the value of AI-involved art, and to whom does it belong? We begin to answer these questions using data from an in-person, juried AI art…
AdaPath: Query-Adaptive Path-Finding via Path-Bank for Multi-Hop Implicit Biomedical KGQA
arXiv:2608.30556v2 Announce Type: replace Abstract: Path-finding over knowledge graphs has become an effective way to ground LLM reasoning on multi-hop questions. However, biomedical QA introduces two distinct challenges that general-domain methods are…
The Infrastructure Divide: Analyzing Community Stance on Data Center Development via Reddit
arXiv:2611.10109v1 Announce Type: new Abstract: The rapid expansion of AI-driven data center construction has outpaced public engagement: an estimated $64 billion in U.S. data center projects were blocked or delayed by local…
META-IR: Meta-Introspective Reasoning for Adaptive Safety Alignment in LLM
arXiv:2611.10089v1 Announce Type: new Abstract: Large Language Models (LLMs) remain vulnerable to jailbreak attacks and harmful-content prompts, particularly under adversarial distribution shifts where static alignment pipelines fail to generalize. Although introspective methods…
AYO: A Safe and Privacy-Preserving Platform for Studying Children's Interaction with LLMs
arXiv:2611.10054v1 Announce Type: new Abstract: Children and adolescents are increasingly interacting with generative AI and large language models (LLMs), yet little is known about how these interactions may influence their development, beliefs,…
LoRazor: Trimming Adapter Execution for Efficient Multi-LoRA Serving
arXiv:2611.10049v1 Announce Type: new Abstract: LoRA offers a parameter-efficient adaptation strategy, supporting multi-user workloads via customized adapters on a shared backbone. However, serving concurrent requests with different adapters introduces substantial computation and…
Subliminal Learning Is Steering Vector Distillation
arXiv:2606.00995v4 Announce Type: replace Abstract: Subliminal learning refers to a student language model acquiring a teacher's traits (e.g. a system-prompted preference for owls) when fine-tuned on the teacher's outputs, despite the outputs…
DeepResearch RewardBench: Evaluating Reward Models for Deep Research Report
arXiv:2611.10156v1 Announce Type: new Abstract: Empowered by Large Language Models (LLMs), deep research systems integrate online information into research reports through multi-step web searching and information synthesis. Evaluating these reports presents a…
OMG-Agent: Toward Robust Missing Modality Generation with Decoupled Coarse-to-Fine Agentic Workflows
arXiv:2602.04144v2 Announce Type: replace Abstract: Data incompleteness severely impedes the reliability of multimodal systems. Existing reconstruction methods face distinct bottlenecks: conventional parametric/generative models are prone to hallucinations due to over-reliance on internal…
LLMs as Judges and Examinees for Medical Item Difficulty Prediction
arXiv:2610.38402v2 Announce Type: replace-cross Abstract: Predicting medical item difficulty from item text alone remains challenging. Prior work has largely relied on linguistic features or text representations derived from general-purpose and domain-specific language…
NoContactNoWorries: Estimating Contact through Vision and Proprioception for In-Hand Dexterous Manipulation
arXiv:2606.24450v2 Announce Type: replace-cross Abstract: Perceiving physical contact is fundamental to dexterous manipulation. While robots often rely on dedicated hardware tactile sensors, humans exhibit a remarkable ability to infer contact by integrating…
ChainFlow: Few-Step Latent Flow Matching for Speculative Decoding
arXiv:2611.08669v1 Announce Type: new Abstract: Speculative decoding has become a standard way to accelerate LLM inference: a lightweight drafter proposes tokens that the target verifies in one parallel pass. Existing drafters face…
Decision Context: Persistent Behavioral Learning for AI Agents
arXiv:2611.08644v1 Announce Type: new Abstract: LLM agents do not learn from corrections across sessions. A user fixes the same mistake at the start of every conversation. Existing agent memory systems primarily handle…
Aspect-wise Pairwise Tournaments for Test-Time Selection
arXiv:2611.08640v1 Announce Type: new Abstract: Scaling test-time compute has proven highly effective at improving LLM reasoning. A popular method is best-of-n sampling, where a model generates many candidate solutions and one is…
Cancellation Is Not Enough: What Retries Leave Behind in an Overloaded LLM Engine
arXiv:2611.08684v1 Announce Type: new Abstract: Clients and proxies in front of an LLM engine retry calls that time out. After a transient overload, these retries can keep the engine from recovering even…
Nemotron-Personas-Finance-Japan: A Synthetic Dataset Capturing Japan's Demographics and Financial Context
Control Room: A central system to track and stop model workloads across our research infrastructure
Learn how OpenAI’s Control Room tracks model workloads across our research infrastructure and can rapidly stop them when safety or security concerns arise.
The brain-as-computer metaphor shaped modern AI but may be misleading the consciousness debate
Consciousness researchers Anil Seth, Ned Block, and Giulio Tononi argue in a Nature commentary that today's AI systems aren't conscious and won't become so on silicon hardware. They consider the assumption that…
AI coding agents still need human cleanup on nearly three out of four tasks in a real-world test
Alibaba's new coding-agent benchmark, Co-Design Bench, tests AI agents on real user requests that already tripped them up in practice. Even GPT-6 Astra, the top performer, fully solves only 26.4 percent of…
AI auditors catch most agent misbehavior but flood reviewers with false alarms, new study finds
AI models can often spot problematic behavior by other AI agents in logs, according to a study by Anthropic researcher Andy Hall. GPT-5 found about 98 percent of reward-hacking cases, but the…
Thinking Machines' Inkling 2 Small drops 316 F-bombs across benchmark runs
Thinking Machines' AI model Inkling 2 Small swears during its internal reasoning, 316 times across 2,655 benchmark runs, according to an analysis by Vals AI. The profanity typically shows up late in…
The gap between open and closed AI models in cyber capabilities keeps shrinking
In the Vals AI benchmark, OpenAI's GPT-6 Sol Daybreak hits 80.8 percent of attack objectives in simulated corporate networks. Worse, the open-weight model DeepSeek V4.1 Flash reaches 72.8 percent at a tenth…
State propaganda in some Chinese AI models is only "a thin layer" that fine-tuning can strip away
Nous Research retrained Alibaba's Qwen3.8-27B to produce Arete-27B, largely stripping out its pro-Chinese bias. The propaganda rate dropped from 22 percent to less than one percent, while general capabilities remained intact in…
Why a new report on Anthropic automating AI development is alarming researchers
Anthropic is using AI to build more powerful AI. And it appears to be working remarkably well, with the company's models improving roughly twice as quickly as a year ago, with little…
Nine AI agents tried autonomous research on open questions from US national labs and none passed the experts' bar
The "Into the Unknown" benchmark tasks nine AI agents with tackling open research questions from U.S. national laboratories. None of the 135 submissions meets all requirements. A total of 114 papers fail…
Google's AI detects high blood pressure from a regular smartwatch without a cuff
Google Research built an AI that spots signs of high blood pressure using only the pulse and motion sensors in a standard smartwatch, no cuff needed. In a 196-person study, its 65.8…
Ripple-65B Preview delivers strong agentic coding results as a fully Apache 2.0 open US model
Ricursive Intelligence is releasing Ripple-65B Preview, a 65-billion-parameter open-weight language model that the company says achieves the top score among open-source U.S. models on the TerminalBench 2.1 coding benchmark at a total…
Hundreds of Artex AI hacking tools found publicly exposed to the web
Security researchers found hundreds of servers hosting the penetration testing tool Artex exposed to the internet, allowing anyone access to the hacking tool.
Frontier Radar #5: If AGI is so close, why is everyone leaving? The neolabs and their bets
OpenAI is talking about the AGI era, and Nvidia CEO Jensen Huang says AGI has arrived. At the same time, co-creators of today's models are leaving OpenAI, Google DeepMind, and Meta to…
METR finds Anthropic may have crossed its own AI acceleration warning threshold back in July
Nonprofit METR spent eleven weeks auditing Anthropic's internal research data and concluded that the pace of model development has roughly doubled, with about 80 percent of the speedup driven by AI-assisted research.…
CZITAPP shows headlines and concise feed summaries, attributes every source and links to the original publisher. Source type is descriptive, not a numerical truth score.